Climate science is, or should be, a branch of experimental physics depending on observation rather than theoretical preconceptions. It is primarily concerned with the relationships between time series of measurable physical quantities and their long term behaviour. Here we develop a rigorous and useful statistical calculus of time series and apply it to real world data. We look at how both global and regional surface air temperature time series are related to atmospheric carbon dioxide concentration. We reveal geographic patterns of climate sensitivity which, to date, have eluded climate modellers.
The full paper can be downloaded here: